Neoadjuvant or preoperative chemotherapy is becoming an important part in breast-cancer treatment and management. In parallel it is important to develop reliable monitoring methods to evaluate the response. Dynamic contrast enhanced MRI (DCE-MRI) has been proven as the most accurate imaging modality to predict residual tumor responses. In this study we will apply serial DCE-MRI studies to evaluate the pre-treatment lesion characteristic features, along with the early response patterns determined in follow-up MRI studies to aim for accurate prediction of pathologic complete remission (pCR), which is the most relevant prognostic factor in patients receiving neoadjuvant chemotherapy. At our institution we have an active on-going neoadjuvant chemotherapy protocol. In previous patient cohorts we have demonstrated that MRI is very helpful to provide response-monitoring information for timely adjustment of regimens. Based on these results we have determined the treatment protocols to be used in the current study. Patients will receive 2 cycles AC treatment (doxorubicin and cyclophosphamide) then followed by 4 cycles Taxane regimen (Nab-paclitaxel with Carboplatin, with trastuzumab for HER2-positive cancer and bevacizumab for HER2-negative cancer). In the current proposal we will take the role of MRI one step further, to investigate whether the pre-treatment tumor characteristics, including morphology, texture, and vascular parameters, as well as metabolic information, can be used to predict whether a patient will achieve pCR. The information that we obtain will reveal the differences between these two groups, and that can be used to build a model to predict the likelihood of achieving pCR. If a patient is expected to show a good response, the protocol can be given as it is. If a patient is expected not to show a good response to reach pCR, a modified protocol should be considered. The response monitoring data may also provide information for modifying the regimen to reach an improved efficacy. The early response after 2 cycles AC, then after addition of 1 cycle of taxane will be evaluated. Overall we will evaluate the accuracy of using pre-treatment MRI characteristic parameters, and early tumor response patterns during the treatment course, to differentiate between patients who achieve pCR vs. those who do not. It is hypothesized that a higher accuracy in prediction of pCR can be achieved using combined pre-treatment lesion characteristics and early tumor response patterns compared to that using either set of data alone.